Sexual behaviour among users of antiretroviral pre-exposure prophylaxis.

The pioneering trials of antiretroviral pre-exposure prophylaxis for HIV prevention have yielded a potent new strategy for the reduction of HIV incidence in atrisk populations. However, pre-exposure prophylaxis is inherently a combination approach: the intervention eff ect will depend not only on drug effi cacy, but also on behaviours of users and providers. Before the fi rst trial fi ndings emerged, scientifi c researchers had already identifi ed the potential for increased sexual risk-taking among pre-exposure prophylaxis users who feel protected from HIV. Concerns about risk compensation behaviour have persisted as the preexposure prophylaxis evidence base has grown. Although fi ndings from trials consistently show reductions in risk behaviour of participants, controlled clinical trials are not an optimum setting to test the association between behaviour of pre-exposure prophylaxis users and their perceptions of drug protection. The study by the Partners PrEP team makes a crucial contribution to the scientifi c literature on risk compensation associated with pre-exposure prophylaxis. This analysis capitalised on the open-label extension of the Partners PrEP trial among serodiscordant heterosexual couples, which was stopped after defi nitive evidence of pre-exposure prophylaxis benefi t was found. HIV-uninfected participants originally randomised to receive tenofovir alone or with emtricitabine continued to take pre-exposure prophylaxis without interruption, and questionnaires tracked sexual risk behaviours 12 months before and after unblinding. Compared with predicted behaviour trends, participants reported no signifi cant change in the frequency of unprotected sex with their HIV-infected partners. Although most sex acts of participants continued to be with their primary HIV-infected partners, investigators also found a signifi cantly increased frequency of unprotected sex and total sex with outside partners over time. However, no corresponding increase in incidence of sexually transmitted infections or pregnancy was reported. This study adds to analyses of behaviour among users of pre-exposure prophylaxis, and acceptability research among populations considering pre-exposure prophylaxis uptake. Empirical data for risk compensation associated with pre-exposure prophylaxis are valuable for several purposes. These data should guide the development of pre-exposure prophylaxis implementation packages for users, which should include culturally tailored behavioural counselling and assessment strategies to minimise risk-taking and to maximise adherence. Results from the Partners PrEP analysis show the need for providers to counsel users on risks both within serodiscordant partnerships and with outside partners. Understanding pre-exposure prophylaxis user behaviour can assist in modelling the eff ect of implementation at the population level, improve cost-eff ectiveness calculations, and help address the structural issues integral to pre-exposure prophylaxis roll-out. Empirical data for risk compensation should also contribute to pre-exposure prophylaxis education for providers. Findings from several studies suggest that providers’ predictions of risk compensation behaviour among patients might aff ect willingness to prescribe pre-exposure prophylaxis. If providers presume that some patients are more likely to increase risk-taking, and therefore should not receive pre-exposure prophylaxis, then anticipated risk compensation could also lead to unequal access. Incorporation of behavioural data into provider training might help to address perceived risk compensation as a deterrent to implementation of preexposure prophylaxis. Open-label extension studies off er opportunities to analyse behaviour among pre-exposure prophylaxis users, but they also present methodological challenges. The notion of preventive misconception suggests that participants in prevention trials often formulate their own perceptions of group assignment and intervention benefi t, which might aff ect behaviour before unblinding. That is, participants who believe they have received an eff ective drug might adjust their risk-taking before knowing their true assignment, which can aff ect analyses based on unblinding. Findings from several trials of pre-exposure prophylaxis show that participants hold beliefs about group assignment, and that perceived pre-exposure prophylaxis effi cacy might increase over time and infl uence behaviour. Future open-label Published Online October 17, 2013 http://dx.doi.org/10.1016/ S1473-3099(13)70251-2